When a small spherical metallic particle is irradiated by light, the oscillating electric field causes the particle's electrons to oscillate coherently. The collective oscillation of these electrons is called the dipole plasmon resonance. The collective oscillation in the electron cloud produces an intense, highly localised electromagnetic field (EMF). In order to produce these collective oscillations, the correct excitation wavelength must be used. This wavelength will depend on both the size and shape of the metal particle being excited. The correct wavelength for a particular particle can easily be determined experimentally or, for certain shapes, may be predicted using known modeling techniques.
It has been shown in [1] that the EMF produced by the plasmon resonance of metal colloids such as gold or silver nanoparticles can be used to excite commonly used fluorescent labels such as Cy5 and Rose Bengal, amongst others. These experiments involved placing gold or silver nanoparticles on a surface and positioning the fluorophores such that they benefit from the enhanced EMF. This results in what is called metal enhanced fluorescence (MEF).
When a fluorophore is placed near a metal particle exhibiting plasmon resonance, the fluorophore benefits from both the localised EMF and from a less-well understood phenomenon in which the fluorescent lifetime is decreased. A similar phenomenon occurs when the Raman scattering of a molecule is examined. Where the molecule if in close proximity to a metallic particle exhibiting localised surface plasmon resonance a dramatic increase in the level of Raman scattering is observed.
As the maximum photon emission of a fluorophore is determined by its quantum yield and fluorescent lifetime, a shorter fluorescent lifetime results in higher photon emission. A fluorophore having a 90% quantum yield and a fluorescent lifetime of 2.5 ns could emit a maximum of 360 million photons/s. In reality, fluorescent dyes can emit several million photons before deteriorating.
By inducing a highly localised enhanced EMF, such as has been demonstrated using gold and silver colloidal arrays, [1,2] as well as in colloidal nanocrystals [3], photonic crystal nanocavities [4], metal nanorings [5] and porous thin metal films [6] and positioning a fluorophore such that it will benefit from it, MEF can be induced and the fluorophore detected, or alternatively Raman scattering detected.
If a device is made such that there is a region in space where only one fluorophore may benefit from MEF at any given time, the position of that fluorophore can be known. By arrangement of a plasmonic structure (e.g. a metallic moiety) around or adjacent to an aperture (pore), a fluorescent particle can be oriented such that it passes through an area of maximum fluorescent enhancement (MaxFE). Suitable pores include: gels (commonly used to confine polynucleotides), organic pores such α-haemolysin, solid state pores such as can be made in silicon nitride, nanoporous materials such as porous gold, colloidal crystals, colloidal cavities, metal toroids, or any other pore less than 10 nm in diameter and preferentially less than 5 nm in diameter, especially between 1.5 and 2.5 nm.
Given the following:                the EMF between two plasmon resonating surfaces increases when the distance between them decreases,        that proximity to a plasmon resonating surface also diminishes the fluorescent lifetime of the fluorophore,        that the distances between these surfaces can be made less than 5 nm,        and that colloids, porous thin films, 2D crystal arrays, and other nanostructures can made less than 2 nm in size,it is reasonable to assume that maximum photon emission can be made to occur in an extremely small region.        
In a similar way, particular properties of a molecule, or regions within a molecule, can be identified by characteristic patterns of Raman scattering. This, as with the identification of the location of a particular fluorophore, the position of a particular feature having an identifiable Raman spectrum can be identified by its proximity to the plasmonic structure.
If an ion is induced to pass through a pore by means of, for example microfluidics or electrophoresis, it can be made to pass through this area of MaxFE such that detection of a single ion-fluorophore conjugate, or a region having a particular Raman spectrum can be detected.
By inducing a fluorescently labelled charged ion such as a protein to pass through a nanopore such that only one ion may travel through the pore at once, quantification of those ions would be possible as follows:                Given two identical fluorescent particle/ion conjugates, as the first passes through the pore and enters the confined EMF, the fluorophore will begin to emit photons. As the fluorophore moves closer to the region of MaxFE, an increasing number of fluorophores are emitted. As photon emission is directly related to the intensity of the EMF, which in turn increases exponentially as the distance to the region of MaxFE decreases, the photon emission will also increase exponentially, producing a photon emission peak in the region of MaxFE and decreasing as the conjugate moves away from the area of MaxFE. The second identical fluorescent particle will produce the same pattern of photon emission. So long as peak photon emission occurs in a discrete region with a thickness less than the distance between the two fluorophore/ion conjugates, the fluorescence produced by these two fluorophores can be differentiated.        Should the two conjugates be spaced so closely together (less than 0.5 nm) that individual detection would be difficult, the presence of more than one fluorophore in the area of MaxFE could be determined by photon emission. Two identical fluorophores exposed to equal EMF strengths should emit roughly twice as many photons/s as a single fluorophore. Given that a region of MaxFE can contain two, three or more conjugates, if the speed at which they travel through the EMF is known, quantification of the number of conjugates which pass through the EMF can be achieved.        
By creating two chambers separated by a barrier which ensures that ions could only pass from one side to the other via a pore, the ions which can be induced to migrate could be quantified regardless of whether the region of MaxFE can contain two, three or more conjugates so long as that number has been determined.
Labelling of Nucleic Acids
Fluorescent labels are commonly incorporated into the both single stranded (ss) and double stranded (ds) polynucleotides DNA and RNA. There are many well-established methods for incorporating fluorescent labels into DNA and by careful choice of both DNA polymerase and fluorescently modified nucleotide, it has been shown [7] that it is possible to fluorescently label all of one or more of the four bases (CGTA). The distances between nucleotides in dsDNA and dsRNA is known to be 0.34 nm, while in ssDNA and ssRNA, the polymers can be in linear conformation, allowing the maximum spatial separation to be greater—although the exact distance varies between 0.5 and 1 nm depending on the stretching force imposed [8].
The Sequencing Process
It is known that precise control of the speed at which a single strand of DNA passes through an organic pore such as a-haemolysin [10] or a solid-state pore in silicon nitride can be achieved [8,9]. By labelling a polynucleotide such that all of one or more of its bases are fluorescently labelled and inducing that polynucleotide by microfluidics, electrophoresis or some other means to pass through an EMF at a known speed such that the nucleotides pass through a region of MaxFE able to contain at least one fluorescently labelled nucleotide, the position of the labelled nucleotides along the DNA fragment can be determined.
A complete polynucleotide sequence could be assembled in the following way:
Assuming all of one base will be labelled at a time, four PCR reactions can be performed in which in each, all of the C, T, G, or A nucleotide bases are fluorescently labelled. The strands are then denatured-forming ssDNA or ssRNA. Each strand is then induced to pass through a nanopore at a known speed and through an area of MaxFE able to contain a known number of nucleotides. As the labelled nucleotides pass through the EMF, their positions can be determined relative to the other labelled nucleotides—providing a distance map for each of the four identical but differently labelled polynucleotides. These four distance maps can then be easily assembled into a complete polynucleotide sequence.
Whole Genome Sequencing
In order to sequence a genome using this method, fluorescently labeled random primers, such as are available for whole genome amplification, could be used. The sequence of these primers would be known and the fluorescent label could serve to indicate the 5′ or 3′ orientation of the DNA fragment as it passes through the pore. The labelled primer would also serve as an initial reference point from which the distance map of the labelled nucleotides could be related. Preferentially, all of more than one of the four bases would be fluorescently labeled and fluorescent labels with different emission spectra would be used to differentiate these bases. More than one fluorescent label spectra could also be used to label all of one type of base providing the labels had distinct max emission spectra.
These DNA fragments could then be induced to pass through the pore as described above at a rate determined by the photon emission. Using the dye Cy5, photon emission rates for MEF of between 3 and 9 million photons/s have been obtained [1]. This could allow DNA to pass through at a rate of slightly less than one million nucleotide bases per second per nanostructure. An array of nanostructures would allow this speed to be multiplied by the size of the array. As these primers are random and the position of the distance maps for the DNA fragments within the genome is not known, whole genome assembly software such as is used for whole shotgun sequencing would be required for accurate alignment and assembly.
Advantages of the Present Invention
The present invention provides a variety of advantages over other methods of DNA sequencing as well as other methods of single molecule detection:                The invention allows enhanced single molecule detection.        The invention can be used to determine the position and speed of single molecules.        Although PCR can be used to fluorescently label single molecules, the present apparatus and method is not reliant on chemical or enzymatic reactions.        The apparatus can be arrayed such that tens, hundreds, thousands or millions of these nanostructures can be in operation at once.        DNA sequencing using the method outlined above would be extremely fast. The cost of sequencing using this method would be limited to the cost of no more than four PCR reagents such as fluorescent labels, nucleotides, polymerases and other minor consumables.Methods of Construction        
The heart of the invention lies in inducing a single molecule to pass through a highly localised EMF in a controlled manner and detecting interaction between the molecule and the EMF. There are many well-researched methods of controlling both proteins and polynucleotides including gels, organic pores, solid state pores, attachment to beads, and immobilisation on a surface. This allows a wide range of devices to be constructed which could control a single molecule such as a protein or a DNA fragment such that it could be made to pass through a highly localised EMF.
The present invention allows any surface to be used so long as the plasmon resonance induced by an incident EMF will induce MEF in a fluorescently labeled single molecule or SERS, and that single molecule can be induced to pass through an area of MaxFE.
The present invention allows any means of inducing a single molecule to pass through a region of MaxFE including microfluidics, electrophoresis, or laser trapping.
Where fluorescence is the chosen method of detection, the present invention allows any fluorescent label to be used which will fluoresce at the frequency of maximum plasmon resonance of the surface being used. Ideally, the fluorescent label will have a high quantum yield, low fluorescent lifetime, low molecular weight, and be easily incorporated by a polymerase. Preferably, the nucleotides themselves are intrinsically fluorescent.
SERS has the advantage that no labeling is required, but rather the specific Raman spectrum of features of the molecule to be identified should be known.
These characteristics allow for the invention to be engineered in at least three general ways, each of which having distinct advantages:                The use of an “organic” pore to control a single molecule allows gold nanoparticles to be covalently bound either to the pore itself or to the surrounding lipids. Alternatively, the organic pore can be placed inside a larger solid state nanopore such as can be made in silicon nitride, and the plasmonic structure (e.g. metallic moiety) positioned such that MEF or SERS is optimized for those molecules or parts of molecules within the organic pore. Alternatively, an organic pore could be placed inside a plasmonic structure such as a photonic nanocavity (see FIGS. 5 and 6). The advantages of this method include the well-characterised behavior of single molecules induced to pass through such a pore, the small and highly reproducible pore diameter an organic pore allows, and the low cost of producing such organic pores.        The use of a solid state nanopore allows for control over the pore diameter and for nanopositioning plasmonic structures such that MEF or SERS is optimised. Ideally photon emission can be further controlled by means of a waveguide. This would allow for more efficient photon detection and higher detection speeds.        A third general method of construction is to allow for a random distribution of highly localised EMFs such that the field intensity at any given point on a porous surface is not known and to allow for the single molecule to pass through this surface at a controlled speed but uncontrolled location (FIG. 7). This has a tremendous advantage over the other two methods in that as neither the location nor the distribution of either the EMFs or the single molecules are precisely controlled, the ease with which such a device can be constructed is correspondingly greater. A method controlling the speed of the single molecules at the intersection of a 20 nm pore and a surface containing randomly distributed EMFs is described below. This method has the advantage in that it is intrinsically multiplexed. Even if the fluorescence produced by a single molecule passing through an EMF is such that only one molecule can be detected per second, the device can be multiplexed such that several hundred or thousand or million pores are in operation at once.Method for Construction of an Apparatus According to the Present Invention and Performing Sequencing        
An apparatus according to the present invention comprises the following:                A first reservoir 19 capable of holding several μL of solution and containing an electrode 26 embedded in its bottom surface;        A layer of silicon (substrate) 11 with an array of nanopores 10, which are 20 nm in diameter and which are spaced 1 μm apart etched in its surface. Electrodes 12 are embedded in the silicon layer 11 such that each 20 nm diameter pore 10 contains an electrode 12, and each electrode 12 is connected to a voltage supply with a high skew rate (not shown);        A porous membrane 22, on which a regular array of gold nanoparticles 14 are deposited, is bonded to the silicon layer. The gold nanoparticles 14 are of known size and geometry, and when excited at the appropriate frequency, produce an EMF;        A gel 24 such as agarose or polyacrylamide is applied to the membrane 22 having an average pore size such that when ssDNA 20 is induced to migrate through the gel 22 it will adopt an approximately linear conformation. This forms the control means in the present embodiment of the invention;        A second reservoir 21 is provided on the other side of the substrate 22 into which a molecule will pass after having travelled through the pore;        An array of voltage modulators (not shown) with a high skew rate are linked to a sensitive photon detector such as an avalanche photodiode (not shown);        A photon detector such as an avalanche photodiode (not shown) is provided;        A means of controlling an array of voltage supplies based on input from a photon detector (not shown) is provided;        A laser of a frequency capable of inducing plasmon resonance in the metal colloids of known size and geometry (not shown) is provided.Method of Operation to Perform Sequencing        
DNA is labelled via PCR as previously described such that all of two bases (C and A or G and T) in a given strand of ssDNA are fluorescently labelled with fluorescent labels which have a maximum emission near, but preferably slightly red shifted in relation to, the plasmon resonant frequency of the metal colloids 14 being used.
The labelled ssDNA is then placed in the first reservoir 19. The membrane 22 coated with a nanoporous gel 24 and containing an approximately even distribution of gold nanoparticles 14 such that the spacing between them varies randomly between 1 nm and 100 nm is placed above the first reservoir 19. The silicon layer 11 is then placed on the membrane 22 and the entire assembly can be immersed in a conducting solution. A positive charge is then given to the electrodes 12 inside the 20 nm pores 10 such that ssDNA 20 will be induced to migrate through the second reservoir 21 into the gel 24, pass through the gel 24 in linear conformation and pass the gold nanoparticle 14 array emitting photons as they pass through the EMF produced by the plasmon resonance. As the fluorescently labelled random primers were used in the PCR reaction, the first photon emission will be produced by either the 5′ or 3′ primer. The photon emission obtained from the primer will immediately indicate the speed at which the ssDNA 20 fragment can be induced to migrate through the EMF and a clear signal still be obtained. As the positioning of the ssDNA 20 in relation to the gold spheres 14 will be different inside each of the 20 nm pores 10 for every ssDNA fragment 20, there will be an optimal voltage which can be applied to each pore 10 at any given time. As outlined above, a distance map containing the distances relative to each other and to the known primers can be created from the photon emission data. This data can then be assembled into a complete genome sequence.
Additional Methods for Preparation of a Substrate and Performing Sequencing
There are many way in which a suitable substrate can be produced using nanotechnological techniques. The following describes two possible ways and the manner in which a sequencing protocol would be preformed.